1. Field of the Invention
The present invention relates to an iterative decoding device and relative decoding method, and more particularly, to an iterative decoding device and related decoding method capable of improving error correction performance.
2. Description of the Prior Art
The communication of information between computers and other electronic devices can be achieved by using a variety of different specifications and technologies. For example, channel coding and error correction are used to reduce error of received signal due to distortion or other interference. Such encoding and error correction techniques can be achieved by using the encoder and decoder at the both end of transmission channel.
For example, any decoding algorithms for low density parity check code (LDPC code), such as belief propagation algorithm, normalized min-sum algorithm (NMSA), and offset min-sum algorithm (OMSA), can be applied to the receiver for decoding the received codeword. In general, an iterative decoder is used to provide decoding results converged to the correct codeword. For example, a message-passing algorithm (MPA) based on soft decision may be used for LDPC decoding at the receiving end. A plurality of received bits (e.g. a plurality of channel results) can be regarded as a plurality of variables representing probability of value “0” or value “1”. The plurality of variables may be represented by a plurality of log likelihood ratios (LLRs) in the decoding algorithm. In the message passing algorithm, the soft information (e.g. LLR) is passed from a bit node to a check node and syndrome of current iteration is calculated. After that, the soft information can be passed from the check node to the bit node and the same procedure is repeated iteratively until the convergence condition is reached (e.g. the check parity of syndrome are all zero). LDPC code is a well-known technique and has been introduced in a large number of materials, such as U.S. published patent application No. 2011/0307755, “STRUCTURED LOW-DENSITY PARITY-CHECK (LDPC) CODE” and U.S. Pat. No. 7,707,479, “Method of generating structured irregular low density parity checkcodes for wireless systems”.
However, for the LPDC decoding, the soft information may be gradually enlarged for improving the reliability of the determination during the iterative decoding process, so that the soft information would exceed the range of finite word length, thus resulting in decoding errors. Therefore, the soft information needs to be scaled to prevent the soft information from exceeding the range of finite word length. For example, please refer to FIG. 1, which illustrates a schematic diagram of a conventional iterative decoding device 10. The iterative decoding device 10 includes a decoder 100 and a scaling unit 102. The decoder 100 is used for receiving a set of soft information Si, iteratively decoding the aforementioned soft information and updating the aforementioned soft information accordingly to generate a set of updated soft information USi. The scaling unit 102 is coupled to decoder 100 for scaling the updated soft information USi to generate a set of scaled soft information SSi acting as an input of the decoder 100 for next iteration in order to avoid the soft information exceeding the range of finite word length during the iterative decoding process. However, the method can only be applied to a regular LDPC decoder, but can not be applied to an irregular LDPC decoder for improving error correction performance of the irregular LDPC decoder. Thus, the prior art has to be improved.